Conventional vehicular cellular radio systems have stringent intermodulation and noise figure requirements, particularly for cell site equipment, where a large number of users may share receiver preamplifiers of a single cell site. An example of a conventional system is the AMPS system such as described by EHRLICH, N., et al., in "Cell Site Hardware," Bell System Technical Journal, January, 1979, pages 153-199. The low noise receiver preamplifier subsystem requirements of the original AMPS system included a 2.5 dB noise figure. In addition, with two RF signals of -35 dBm each at the input of the receiver preamplifier, the third-order intermodulation product at the output should be greater than 65 dB down from the level of each of the two input RF signals. This is equivalent to an input third-order intercept of -2.5 dBm, and translates to a spurious-free dynamic range of about 83 dB, assuming a channel bandwidth of 30 kHz.
Modern cell site receiver systems are capable of operating in a vehicular cellular environment characterized by high user density with all available channels loaded during the busy hour for some cells, and thus require cell site receiver preamplifiers and receivers which have high dynamic ranges. As cellular systems gradually adjust to using the U.S. digital cellular standard (US TIA TDMA; Telecommunications Industry Association Time Division Multiple Access, also commonly referred to as Digital AMPS), the overall dynamic range requirements are expected to remain stringent.
The degradation of system performance due to third-order intermodulation effects typically occurs when strong and weak signals are simultaneously amplified by the same preamplifier. Intermodulation products resulting from stronger signals may fall on or near those frequencies at which weaker signals are being received, effectively producing co-channel interference in the desired weak signal. This is particularly likely in cellular systems where the channel spacing among voice channels used in a given cell is uniform.
The input signals particularly handled by a cell site preamplifier include signals received from vehicles located near the base of a tower or building on which the receiving antenna is mounted, as well as signals received from vehicles located several miles from the cell site. Typically, cell site antennas have radiation patterns which do not provide significant antenna gain for locations at ground level and in the immediate vicinity around the cell site. Therefore, signals from vehicles located near the base of the receiving antenna typically do not produce the highest signal levels processed by a cell site preamplifier. Instead, the highest input signal levels typically come from nearby vehicles and portable terminals within the main coverage area of the antenna, such as would be the case for the fourth floor of an adjacent building. The closest distance from a cell site antenna to the nearest user terminal is typically several hundred feet.
Optical fiber cable transmission links have been used to provide low-loss transmission of RF and microwave signals over reasonably long distances. Apparently, the dominant sources of non-linearity (which causes intermodulation distortion) for such systems are optical components such as laser or light-emitting diodes used in converting electrical signals to light signals.
Many of the applications for transmission of satellite down link signals do not require extremely wide overall dynamic ranges for the optical link. Most satellite RF transmission links are ideal candidates for use with fiber-remoting technology, since satellite links are typically not affected by rapid deep fading and do not need to handle a wide range of incoming signal levels from one or more satellites simultaneously. On the other hand, cellular receiving systems must typically handle a wide range of signal levels simultaneously.
The overall dynamic range requirements are even more stringent for possible future systems which may employ low power (e.g., +10 dBm output) hand-held transceivers in a micro-cell environment. This is essentially a result of the potential for the remote antenna to simultaneously receive signals from high-powered vehicular terminals (transmitting up to +35 dBm ERP effective radiated power) and hand-held portable terminals (transmitting up to +28 dBm ERP) near the antenna, while simultaneously receiving a weaker signal from a hand-held +10 dBm output portable located 1500 feet from the antenna and perhaps not directly in a line of sight path.
In another potential scenario, an omni-directional remote antenna unit could be located on a street corner and simultaneously receive two signals from vehicles located 30 feet from the antenna, both transmitting to distant vehicular cellular cell site antennas. These vehicular terminals could be either accessing vehicular cell sites associated with the same system or could be operating in the extended spectrum segment assigned to the alternate cellular carrier and located adjacent to the frequencies utilized by the micro-cell. In this case, the received level at he micro-cell could be as high as -15 dBm for each signal. If, instead, there were two nearby hand-held portables equipped with reverse channel power control, each with an output power of +8 dBm ERP and located 15 feet from the micro-cell, the received level could be as high as -35 dBm for each signal. On the other hand, the desired input signal level coming from a distant hand-held portable user could be about -110 dBm. Based on published propagation models, the portable could actually be located anywhere from 200 to 1000 feet from the antenna. For example, see BERNHARDT, R. C., "The Effect of Path Loss Models on the Simulated Performance of Portable Radio Systems," Proceedings of Globeco '89, November 1989, pages 1356-1360.
Therefore, a spurious-free dynamic input range of greater than about 75 dB is required where two hand-held portables generate intermodulation products, and greater than about 95 dB is required where the two nearby vehicular transmitters are causing intermodulation distortion. For the case where an FM system's performance is determined by the level of co-channel intermodulation noise, the desired incoming or received signal level must be greater than the level of the intermodulation signal by the magnitude of the capture ratio, which is typically about 6 dB.
Many cellular systems employ a limited range adaptive power control for mobile transmitters to ideally utilize the minimum output power required to maintain communication with the cell sites. The typical range of adjustment by the adaptive power control is 28 dB total, with discrete steps approximately 4 dB apart. The primary benefit is reduction of co-channel interference, since the same channels are typically re-used in other vehicular cells and micro-cells within the same local cellular system. A secondary benefit of power control is the overall reduction of the incoming signals and the overall input signal level produced thereby at the cell site, thus easing pre-receiver preamplifier dynamic range requirements.
In the case of the micro-cell environment, transmitters with variable output power under cellular system control are also utilized, and these transmitters would ideally utilize a range of maximum output power levels and range of power adjustments. The reason is that while present hand-held portable units have a maximum output power of 0.6 watts, the apparent demand for a lower battery drain, shirt-pocket size unit is expected to create a market need for lower output power units capable of operating in a micro-cell environment. Also, the mobility of the terminal user allows the transmitter to be located quite close to the micro-cell antenna units. Typically, micro-cell antenna units may be placed within 10 to 30 feet of ground level.
While it is likely impractical or not cost-effective to optimize fiber optic cable antenna-remoting systems to operate in the most difficult signal level range environments, a reasonable level of performance comparable to that obtained for present cell site receiving systems, e.g., equivalent to nearly 85 dB of usable dynamic input range, would be acceptable for many signal level environments.
Early antenna-remoting systems employing intensitymodulated fiber optic cable transmission were designed to obtain a spurious-free dynamic range of 53 to about 73 dB. For example, see MEYER, L. J., "Using Fiber Optic With Analog RF Signals," Proceedings of VTC '89, May 1989, pages 398-400. Recently, a spurious-free dynamic range of about 77 dB has been reported by FYE, D. M., "Design of Fiber Optic Antenna Remoting Links for Cellular Radio Applications," Proceedings of VTC '90, May 1990, pages 622-625.
Linearization techniques have been utilized for laser diodes in laboratory settings to further provide reduction of intermodulation products generated by laser diodes. However, such linearization techniques are costly and complex. Such techniques include active pre-distortion, electro-optical feed-forward, and electro-optical feedback configurations. These active linearization techniques for laser diodes, such as disclosed by DARCIE, T. E. and BODEEP, G. E., "Light Wave Subcarrier CATV Transmission Systems," IEEE Transactions on Micro-wave Theory and Techniques, Vol. 38, No. 5, May 1990, pages 524-533, are apparently not yet feasible for widespread deployment. Due to the reported existence of frequency-dependent and frequency-independent intermodulation processes, the linearization alternative, especially for wide band RF signal transmission, appears impractical. Second-order predistortion has also been employed for transmitting two analog video signals or a 960 voice channel FDM (Frequency Division Multiplexed) noise-loaded test signal over a fiber optic cable utilizing a light-emitting diode, as reported by RAMADAN, M., "Analog Signals Transmission over Optical Fiber Systems," 1985 IEEE MTT-S Digest, June 1985, pages 303-306; but nonetheless, third-order distortion products were not cancelled.
Two antenna remoting systems which do not employ active linearization techniques to cancel laser diode distortion are described by FYE, D. M., "Design of Fiber Optic Antenna Remoting Links for Cellular Radio Applications," Proceedings of VTC '90, May 1990, pages 622-625 and by MEYER, L. J., "Using Fiber Optics With Analog RF Signals," Proceedings of VTC '89, May 1989, pages 398-400. These systems were apparently designed to provide adequate cost-effective performance in the particular environment in which they were required to be operated.
These two systems both use fixed gain RF preamplifiers, the gains of which were set so that the highest expected incoming signal would not over-drive the laser diode utilized by the optical transmission link and cause significant intermodulation noise. Neither of these two systems described in the literature was shown to have an 85 dB usable dynamic input range. It is noted that for a typical cell site application, the fiber optic link output feeds a receiver multi-coupler having one or more integral low-noise amplifiers with an optimum signal-to-noise ratio and signal-to-intermodulation noise ratio performance, so as to provide acceptable signal levels to a receiving system.
In one of the above-mentioned systems, described by Meyer of DECIBEL, the FP laser modulator has an input third-order intercept of +24 dBm. FIG. 1 shows a block diagram of an antenna remoting system described by Fye of GTE. In the GTE system, preamplifier 10 is provided having a noise figure of 2.7 dB and an RF gain of 37 dB. Further, a fiber-optic transmission link 12 is provided having an RF loss of 27 dB. An equivalent amplifier 14 shown in FIG. 1 represents the overall equivalent net gain of the GTE system, which is 10 dB. In the GTE system, the DFB laser modulator apparently has an input third-order intercept of about +27 dBm. The link output noise (due principally to the laser noise) is about -115.5 dBm measured in a 30 kHz band width, with no RF input signal. Therefore, if an RF signal at -125.5 dBm were connected to preamplifier input 16 as shown in FIG. 1, the output response of the system would be about equal to the noise level present at the output of the link with no input signal present at preamplifier input 16. A 37 dB RF gain and 2.7 dB noise figure for preamplifier 10 appear to be well suited for microcell applications, if only the lower end of the preamplifier input dynamic range is considered.
If a composite RF signal comprising two individual signals at -48.5 dBm each were connected to the preamplifier input 16, the level of the resulting two-tone third-order intermodulation product at link output 18 would still be -115.5 dBm, which is equal to the level of the random noise output for a 30 kHz bandwidth when no RF input signal is present. For the system, the input signal range over which the output third-order intermodulation product is at or below the output level of random noise with no input is known as the spurious-free dynamic input range, and for this case is 77 dB, determined by subtracting 48.5 from 125.5. This example also indicates that the link performance may be degraded if two or more input signals having levels greater than -48.5 dBm were connected to preamplifier input 16. Since the preamplifier gain is 37 dB, this translates to a maximum allowable level at laser diode modulator input 20 of -11.5 dBm for each of two signals, giving a two-tone third-order intermodulation product of 77 dB below the level of each of the two high level signals. The composite power at laser diode modulator input 20 for these two signals would be -8.5 dBm.
It is worthwhile to consider the effect of connecting signals to preamplifier input 16 at levels higher than -48.5 dBm each, particularly since it is possible in actual applications for a nearby mobile or portable transmitter to cause an input as high as -15 dBm. In a typical scenario, however, two signals at -35 dBm each are injected at the preamplifier input 16. In this case, the result for the GTE system would be: (a) two high-level signals would appear at transmission link output 18 at a level of -25 dBm each; and (b) two signals due to the third-order intermodulation of the two high-level input signals would be generated within the system and would appear at output 18 of link 12, at about -50 dB below each fundamental signal, corresponding to a power of -75 dBm for each intermodulation product at link output 18, which is equivalent to -85 dBm each referred to preamplifier input 16. These two extra undesired signals may unfortunately fall on the same frequency as a desired voice channel signal, and thus the level of a desired incoming signal falling on the same frequency must, due to capture effect, be at least -79 dBm or greater at preamplifier input 16 in order to capture the receiver at the cell site. Thus, when two higher level input signals at -35 dBm each are input at preamplifier input 16, the effective spurious-free dynamic input range for other input signals is reduced to about 44 dB (about 79-35, which is about 44 dB).
Consequently, using the FIG. 1 system, if a hand-held portable, which may be at some distance from the micro-cell antenna, were only injecting a -110 dBm signal (as was the case for a previous example), the call in progress from the portable would likely be completely disrupted by the intermodulation signals if one intermodulation product fell within the same voice channel as that occupied by the desired -110 dBm input signal. This is obviously quite undesirable.
If two input signals at an even higher level, e.g., -31 dBm for each signal, were instead injected, the fundamental signals at link output 18 would be -21 dBm each, and two-tone third-order intermodulation products would be about 42 dB below the fundamental signals, corresponding to -73 dBm for each intermodulation product referred back to the level at preamplifier input 16, effectively interfering with possible desired input signals at preamplifier input 16. In this case, for an input signal to overcome the intermodulation noise and produce a usable demodulated signal at the FM receiver, it would have to be at a level of at least -67 dBm.
Systems are typically provided to allow system control of output power for mobile and portable terminals to ease the overall dynamic range requirements for micro-cells and antenna remoting links. Unfortunately, as shown in the example above, the benefits provided by controlling the mobile or portable transmitter power would likely not be adequate to provide intermodulation performance for a micro-cell antenna remoting link, which is comparable to that now being offered by conventional cell site equipment. This is particularly true when a nearby transmitter is not controlled by the remote micro-cell antenna of interest, such as when the nearby transmitter is either served by a large vehicular cell or another local cellular system. Another system has been proposed to resolve the high-level signal problem as discussed above by substituting the preamplifier of a remoting system with an AGC (Automatic Gain Controlled) amplifier. See HOWAT, F., "Cell Like Performance Using the Remotely Controlled Cellular Transmitter," Proceedings VTC '89, May 1989, pages 535-541. For a system using a laser diode, the AGC action would ideally limit the output power of the preamplifier, thus preventing high total power levels from over-driving the laser diode. The gain of an AGC amplifier may alternatively be determined by the average of the signal envelope or by the peak thereof. However, this particular remedy also has a number of undesirable effects in resolving the high-level signal problem, which are noted as follows:
(a) If the AGC time constant is too slow, high signal levels may over-drive the laser diode until the loop of the AGC responds. Since the gain of the AGC is highest when the input signal level is low, the first parts of an incoming signal are amplified at maximum gain until the automatic control takes over. Consequently, a sharp peak, or overshoot, may appear at the output of a remoting system utilizing an input AGC preamplifier. PA1 (b) If the AGC time constant is too fast, information transmitted through the use of AM (Amplitude Modulation) signals may be removed due to varied amplification by the AGC amplifier. Also, because of intentional AM on one or more high-level signals, undesirable AM may be impressed on low-level signals without intentional AM. Moreover, wide-band feedback loops typically are more prone to exhibit instabilities. Received TDMA (Time Division Multiple Access) bursts may also cause system performance problems. PA1 (c) If the AGC loop stability parameters are not properly defined, the AGC amplifier output may exhibit undesirable ringing, overshoot, or undershoot. This is particular true when sudden large input level changes or transients are encountered. PA1 (d) If the AGC amplifier has excessive AM/PM (Amplitude Modulation/Phase Modulation), the PM generated by the amplifier may degrade the quality of the signals being amplified. PA1 (e) Depending on the specific implementation, AGC amplifiers may exhibit poor noise figure and/or degraded linearity (including excessive compression, expansion or third-order intermodulation effects) at some input levels. PA1 (f) Since most cellular systems employ reverse channel power control, the effect of an AGC amplifier would be to remove the desired effect resulting from a change of transmitter power level. This reduction in overall link gain as the system input power level increases may in turn cause a system performance degradation or transmitter power control instability.
Some of the problems listed above arise as a result of the fact that an AGC amplifier is typically a type of feedback amplifier, which has inherent technical challenges regarding loop timing. For example, a correction is only initiated when it is already "too late," and an error (in the sense of an unduly high or low instantaneous output level) has already been detected at the output.
Another way some systems have attempted to resolve the high-level input problem is by selecting a lower fixed RF gain for the preamplifier, or by installing a fixed attenuator in series with the input and/or output of the preamplifier. These measures effectively reduce the signal-to-noise ratio at the laser diode for the entire input range, and potentially degrade the overall performance at all input signal levels as a trade-off for providing lower intermodulation distortion levels for high system input power conditions.
Another way of addressing the high-level input problem is to utilize a limiting amplifier or a fixed gain amplifier with a low output power capability. In most cases, this approach increases the level of intermodulation products generated by the preamplifier and generally raises the level of intermodulation noise generated by the preamplifier, making it approximately equivalent to the level of the intermodulation products caused by the laser diode or light-emitting diode, and typically doubling the total transmission link intermodulation noise. There is also a large potential for introducing AM/AM distortion and AM/PM distortion for the system input signals being processed with this approach. Without employing some type of output power compression or limiting capability, the preamplifier would be capable of over-driving the laser or light-emitting diode by about 20 dB or more, potentially introducing a substantial amount of intermodulation noise.
It should be noted that the above-mentioned high-level signal problem may also occur when using free-space optical links, since the laser or light-emitting diode noise and intermodulation characteristics thereof could also potentially limit the overall dynamic range of an RF transmission link using free-space optical transmission.
A further potential remedy for the above-mentioned problem would involve the use of advanced antennas, such as those adaptive phased arrays capable of attenuating signals with too high an input signal level. An additional possibility is the use of a fixed array with reduced capability to receive signals from transmitters located extremely close to and underneath the antenna, for example, if it were mounted to a tall pole or antenna mast.
An additionally related problem which may potentially occur with micro-cell antenna remoting systems relates to the fiber-optic cable and/or free-space optical link used in transmission of signals in the opposite direction, that is, from the cell site toward the remote antenna unit. In the event a single cell site transmitter's power control function misadjusts the output power to an abnormally high level, the laser or light-emitting diode of the optical transmission link may generate excessive intermodulation products and cause the remote unit to no longer meet FCC requirements for spurious radiation. The potential also exists for causing interference to other radio systems using radio channels adjacent to those utilized by the micro-cellular system employing the antenna remoting system. To address this problem, many of the remedies listed above would likely not offer a practical solution.
A technique which is employed in voice processing systems (including those used in cellular networks and satellite transmission systems) is known as companding. This involves a device located at the source end (a compressor) which compresses the dynamic range of a given signal (of a typically low band width) by a given factor before it is transmitted over a limited dynamic range transmission path. At the distant end, a device known as an expander restores the original dynamic range of the complex signal, utilizing the same factor as utilized in the compression process. Such a device is of limited value for most micro-cell antenna remoting applications since it would need to have an extremely wide loop band width (likely about 50 MHz for this application) in order to instantaneously compress and then expand the dynamic range. Further, it is noted that the compressor and expander devices of a companding system both utilize non-linear feedback loops (potentially exhibiting transient response errors and instability) to vary the gain as a function of the instantaneous input or output level, potentially causing undesirable distortion effects due to residual expansion or compression.
The technical literature apparently does not address means typically provided for alarm reporting, self-test, and remote control functions, for fiber optic cable or free-space optical antenna remoting systems. Such systems may employ separate means such as on-site troubleshooting or perhaps direct wired or dial-up alarm reporting facilities. One system described in the literature utilizes a fiber optic cable system to distribute both microwave signals and digital base band signals which are both directly intensity modulated. See HEIM, P. J. and McCLAY, C. P., "Frequency Division Multiplex Microwave and Baseband Digital Optical Fiber Link for Phased Array Antennas," IEEE Transactions on Microwave Theory and Techniques, Vol. 38, No. 5, May 1990, pages 494-500.
One disadvantage of utilizing digital modulation is the inherent phase noise build-up on all microwave signals resulting from the dual-mode (RF and Digital) direct modulation process. As a result, some types of antenna remoting system applications would not permit the use of direct baseband digital modulation for control and alarm reporting functions.
Due to the limited effective dynamic range of existing remote micro-cell antenna units or radio ports, a relatively large quantity of remote radio ports would be required using such existing systems for micro-cell applications in a given geographical area. Thus, a significant amount of hardware and maintenance costs for additional remote radio ports would result if existing systems were used.
The high dynamic range required for a micro-cell radio port has recently been discussed in detail. A radio port with a high dynamic range will potentially provide a greater practical operating distance for many micro-cell applications, so as to potentially reduce the number of radio ports required to cover a given geographical area or indoor facility. If radio ports with low dynamic range are employed, the number of radio ports required to cover a given area would be higher, likely increasing the system cost.
In contrast to the above described existing systems, the dynamic range enhancing system, and the process for increasing the dynamic range of a transmission link, according to the present invention, overcome most if not all of the above-mentioned problems and result in a dynamic range enhancing system which provides a number of advantages.